Committee reviews APD automated license-plate reader policy; staff stress verification and 14-day retention
Loading...
Summary
Anchorage Police Department staff presented a dash-camera-based automated license-plate reader policy using Axon systems, emphasizing two-step verification, a 14-day automatic retention period unless data are entered into evidence, and prohibitions on uses tied to protected status; committee members pressed on AI, body-cam use and oversight.
Sean Case of the Anchorage Police Department presented the department's proposed automated license-plate reader (ALPR) policy and told the Public Health and Safety Committee the system will be integrated with Axon dash cameras and run from patrol vehicles rather than fixed roadside cameras. Case said APD intends to load narrowly defined criteria into the system such as stolen vehicles, violent-person warrants and Silver/Amber alerts so the system only generates hits on specified priorities.
Case described a two-step verification process when the system flags a plate: an officer must verify the hit against the state system and visually confirm that the license plate corresponds to the vehicle in front of them. He emphasized that an ALPR hit alone does not justify a stop; officers must still articulate reasonable suspicion or probable cause grounded in physical descriptors, proximity to a crime or other articulable facts before initiating a traffic stop.
The policy would limit use of ALPR data for harassment or personal use and bar targeting based on protected characteristics. Case said the department will require training before granting access and will log, audit and encrypt access to the system. He described a 14-day automatic retention window for dash-camera ALPR reads; data not pulled into the evidence system are purged after two weeks, while items tied to an investigation (for example, homicide evidence) follow longer statutory retention rules.
Committee members asked whether APD is using Axon (Case confirmed Axon is the vendor, not Flock Safety), whether the product uses artificial intelligence and how protected characteristics apply when information appears on a vehicle (for example, bumper stickers). Case said the product has object-recognition features that could be characterized as AI for tasks such as recognizing visible damage on a vehicle, but he distinguished that from generative AI. He also said the policy prohibits using ALPR to broadly query plates tied to attendance at protected activities (for example, a protest) and described quarterly privacy reviews with the municipal attorney's office and an annual audit report to monitor impacts.
Public comment later in the meeting raised concerns that ALPR text outputs can be queried or combined with other analytic tools and urged civilian oversight of retention and access policies. APD staff said the system will not initially be configured for long-term tracking locates and that flags are entered manually and can be given time limits, but acknowledged follow-up questions remain about technical and oversight details.
The committee did not vote on the policy at this meeting; members requested follow-up on select technical and oversight questions.
